The Condor 96:106&1075
8 The Cooper Ornithological
Society 1994
THE EFFECT OF FOOD AVAILABILITY
ON TIME AND ENERGY
EXPENDITURES
OF TERRITORIAL
AND NON-TERRITORIAL
HUMMINGBIRDS
DONALD R. POWERSAND TODD MCKEE
BiologyDepartment, GeorgeFox College,Newberg,OR 97132
Abstract. We studied the time and energy allocations related to territorial behavior in
male Blue-throated Hummingbirds (Lampornis clemenciae;about 8.3 g) under conditions
of unlimited and restricted food availability. When food was unlimited. territorial males
avoided inter-specific aggression,chasing only 11% of the inter-specific’intruders (Blackchinned Hummingbirds, Archilochusalexandri;about 3.5 g). Thus, when food wasunlimited,
inter-specific intruders were able to forage efficiently, meeting their estimated daily energy
requirement with ease(27 Id/day). Conversely, 8 1% of intraspecific intruders were chased,
and intra-specific intruders were able to feed at territorial feedersonly when the territorial
male was away. Chases of intra-specific intruders were longer and appeared to be more
intense than chasesof inter-specific intruders. When food was restrictedterritorial activity,
including the total number of chasesengagedin by the territory owner, was significantly
reduced,although the basic characteristicsof territorial behavior (e.g., chasesand displays)
did not change.Territory ownerschaseda higher proportion of inter-specificintruderswhen
food was restricted (48%), suggestingan increasein inter-specific competition. A high proportion of intra-specificcompetitorscontinuedto be chased(80%), althoughthe total number
of intra-specific intruders was lower. We believe that variations in the cost of territoriality
are dependent primarily on the level of intra-specific competition. This is supportedby the
fact that when food was restricted to an amount that could support a maximum of 1.4 L.
clemenciae(based
on doubly labeledwater measurementsoffield metabolic rate in a previous
study), energy intake by the territory owner decreasedfrom 114 kJ/day to 64 Id/day, with
the primary difference being number of intra-specific chases.These data also suggestthat
the exclusion of other hummingbird species might be based strictly on the amount of
available food (energy). When food is restricted, inter-specific competition is more costly
to the territory owner causingthe exclusionof a higher proportion of inter-specificintruders.
The high proportion of intra-specific intruders that are chasedin either experimental condition suggeststhat territorial behavior in L. clemenciaemight have functions other than
resourceprotection per se, such as social functions related to their mating system.
Key words: Archilochusalexandri,Lampomis clemenciae;Trochilidae;territoriality;food
resourcelimitation.
INTRODUCTION
Animals rarely exist under conditions of unlimited food (energy) availability. Becausean adequate supply of food is vital for survival, animals
often exhibit behaviors (frequently involving aggression) that insure accessto food in a given
area. An example of such behavior is territoriality, which is presumed to be adaptive when the
benefits of exclusive use of an area or other resource (e.g., food) exceed the cost of defense
(Brown 1964, Carpenter and MacMillen 1976).
This economic analysis of territorial behavior is
largely basedon studiesof feeding territories (territories defended during the non-breeding sea’ Received6 April 1994. Accepted 18 July 1994.
son) becausetheir characteristicsare presumably
tied to the density and distribution of a specific
resource, primarily food (e.g., Carpenter and
MacMillen 1976, Kodric-Brown and Brown
1978). However, many animals that exhibit territoriality defend resourcesother than food and
in some casesappear to tolerate, at least over the
short-term, an energy deficit (e.g., trout and pupfish during the breeding season;Feldmeth 1983).
It is also important to note that, in territorial
systemswhere this occurs,the importance of behaviors that are prioritized above energy acquisition is not always obvious. For example, Powers and Conley (1994) in a study of Blue-throated
Hummingbirds (Lampornis clemenciue) concluded that although territorial males exhibited
the classicbehavior of hummingbirds defending
[1064]
HUMMINGBIRD
feeding territories, the dynamics of their territorial system was likely driven to a large degree
by intra-specific social interaction. Their conclusions were basedon the fact that territorial males
did not defend their food source efficiently and
becausestrong aggressionwas exhibited only towards conspecifics. Examples such as this underscore the importance of not assuming a territorial system is strictly energy based becausea
food sourceis defended. Rather, an understanding of territorial dynamics should be based on a
complete analysisof the important behaviors and
the energeticsof the organism involved.
Hummingbird energeticshas been of interest
to biologists for some time, as evidenced by the
many papers published on the subject. A large
proportion of these studies have included estimates ofthe energycostsassociatedwith foraging
(e.g., Ewald and Carpenter 1978, Wolf 1978)
with the primary focus being on territorial species. To guarantee a high energy intake, many
hummingbirds forageat and defend quality food
sources(i.e., feeding territories), which involves
making frequent short flights to feed, chasingintruders, and performance of aggressivedisplays
(e.g., Stiles 197 1, Kodric-Brown and Brown
1978). Other hummingbird species,however, are
non-territorial, and rarely engage in aggressive
defense. Examples of non-territorial hummingbirds include subordinate species that are excluded from food sources(e.g.,Pimm et al. 1985)
and traplining species that forage over a wide
area (Feinsinger 1986). These speciesundoubtedly work hard to meet their energy demands
because of higher foraging costs (Wolf 1978).
Territorial species are easier to work with because their activities are generally confined to a
specific area, whereas non-territorial hummingbirds can be widely scattered and less predictable. Thus, few empirical data are available on
non-territorial hummingbirds, which limits our
understanding of their energeticsand makes direct comparisonswith territorial speciesdifficult.
Comparisons that might be attempted between
territorial and non-territorial hummingbirds are
also complicated by differences in the thermal
environment (e.g., conductive, convective, and
radiative heat transfer) experienced by individual specieswhich can have a significant impact
on field metabolic rate (FMR) (e.g.,Bakken 1976).
It is becoming increasingly clear that hummingbird behavior is influenced by the distribution of food (energy) resources.For example,
TERRITORIALITY
1065
breeding male Anna’s Hummingbirds (Culypte
anna) and Calliope Hummingbirds (Stellulu cufliope) modified their foraging behavior as food
distribution changed,yet they kept other aspects
of their behavior constant (Armstrong 1987,
Powers 1987). These behavioral changesinclude
aggressiveactivities against competitors as well
as direct foragingactivities. Thus, changesin food
quality and distribution impact behaviors that
contribute to FMR. This might explain Powers
and Conley’s (1994) observations of L. clemenciue and the Black-chinned Hummingbirds (Archilochusulexundri) during conditions of unlimited food availability, where L. clemenciae
essentially abandoned inter-specific territoriality, and focusedsolely on conspecificinteractions.
Because of this A. ulexundri, a non-territorial
forager where they coexistedwith L. clemenciue,
was in some casesable to remove more energy
from territorial feedersthan the territory owners.
In addition, doubly labeled water measurements
suggestthat under these conditions A. ulexundri
had a much lower FMR than the larger L. clemenciue.Archilochusulexundriwasthereforeable
to operate at a higher energetic efficiency as a
non-territorial forager than the territorial L. clemenciue.
The next logical step in examining the relative
cost of territorial vs. non-territorial foraging in
hummingbirds is to determine if changesin foraging behavior and relative foraging costsoccur
when food supplies are restricted. To test the
effects of energy restriction, we used the sympatric populations of L. clemenciue and A. ulexundri described above. We hypothesize two
possible outcomes: (1) L. clemenciue will continue to be aggressiveagainst conspecificintruders, but will exclude a higher proportion of interspecific intruders, or (2) both specieswill alter
their foraging behavior in order to adapt to new
levels of food availability (e.g., Stiles 1971 vs.
Powers 1987). If the first hypothesis is correct,
then territoriality with regard to conspecificsis
obligatoryfor L. clemenciueandprobably evolved
in responseto selective pressuresother than just
the need to protect a food source (e.g., position
in a lek, seeKuban and Neil1 1980). In this case,
A. ulexundri should be forced to forage in other
areas (presumably at a higher cost) as suggested
by Pimm et al. (1985) and their densities in the
study area should decrease. If the second hypothesis is correct, then the dynamics of the L.
clemenciue territorial system should be similar
1066
DONALD R. POWERS ANDTODD MCKEE
to that described by standard cost-benefit models. In this case foraging costs of both species
would likely be higher due to the increased cost
of finding food. Determining which of the above
hypotheses are correct will provide insight into
the evolution of territoriality in hummingbirds
and the degree to which these species can adapt
to different or changing environments.
METHODS
STUDY AREA
We conducted our study during June 199 1 at the
American Museum of Natural History’s Southwestern Research Station in the Chiricahua
Mountains, Cochise County, Arizona (latitude
3 1”50’N, longitude 109”15’W, altitude 1,700 m).
The riparian habitat surrounding the station is
bordered by oak woodland and a mixed deciduous/coniferous forest. For a more complete description see Pimm (1978). Small insects of a size
presumably suitable as hummingbird prey were
abundant during the study.
EXPERIMENTAL
PROTOCOL
Four locations at the Southwestern Research Station (SWRS) were supplied with a feeder around
which birds could establish territories. Feeders
were located approximately 100 m apart. Each
feeder consisted of four 12-mL syringes (Monoject #5 12910) with regular Luer slip tips. The
tips of the syringes were cut off to slightly enlarge
the hole and painted red with nail polish. Syringes were filled with 20% sucrose solution (1 .O
g sucrose mixed with 4.0 g water). Syringes were
inserted through a Plexiglas@ plate suspended
from an aluminum pole.
Two conditions of food (sucrose solution)
availability were used: (1) unlimited, i.e., food
was always available, and (2) restricted, where
each territory was provided with only 32 ml of
sucrose solution per day, an amount equal to
about 114 W/day (assuming 0.2 16 g sucrose/ml
and 16.5 kJ/g sucrose; Weast et al. 1983). The
amount of energy provided in the latter condition exceeded the energy needs of a single L.
clemenciae (82 kJ/day based on doubly labeled
water measurements; Powers and Conley 1994)
by 39%, which provided opportunity for successful intrusion. When food was restricted 20
ml (71 kJ) of sucrose solution was added to the
feeders at 04:30 hr, which was before the birds
became active, and 12 ml (43 kJ) at 15:OO hr.
Providing nectar in this manner more closely
mimics nectar production in natural flowers,
which is often somewhat bimodal, with nectar
production high in the morning and a second
smaller peak in the afternoon (e.g., Stiles 1975).
During the first half of this study all territories
were food restricted whereas all territories had
unlimited food during the second half of the study.
Energy and time costs during the active period
for both the territory owner (L. clemenciae)and
intruders (A. alexandri), and energy resource
management by the territory owner, were evaluated using time-budget analysis, territorial food
depletion measurements, and measurements of
feeding rate. The hummingbirds’ active period
was divided into three parts: morning (05:OO hr
to 09:OO hr), midday (1O:OOhr to 15:OOhr), and
evening (16:OO hr to 19:OOhr).
TIME BUDGETS
We measured time budgets on individual territorial L. clemenciaemales defending each of the
four experimental feeders and intruding A. alexandri. All territorial L. clemenciaeand many
of the intruders could be identified by color
markings on their backs (see Powers and Conley
1994). Because we could not identify each of the
intruders, the impact of individual intruders could
not be assessed. Fifty hours of time budget per
species were obtained during each experimental
manipulation of food availability. Time budgets
were assessed by positioning an observer approximately 20 m from the territorial feeder, with
frequency and duration of specific activities recorded for both species. Activities measured for
L. clemenciae were perching, nectar feeding,
chasing, miscellaneous flight, and out-of-sight
time. Chasing was subdivided into two categories: intra-specific and inter-specific. Miscellaneous flight included flights around the territory
not associated with feeding or chasing. Flycatching was also included in miscellaneous flight because it constituted only a small portion of L.
clemenciae’s daily activity. Out-of-sight time
(00s) included periods when the territory owner
could not be seen by the observer. Time budget
data for L. clemenciaewere recorded with a TRS
100 lap-top computer (Tandy Corp.). Activities
recorded for A. alexandri were nectar feeding and
chasing. Timed activities for A. alexandri were
recorded using stop watches. Time budgets were
recorded between 05:OO hr and 19:00 hr during
1-hr observation periods. Schedules were deter-
HUMMINGBIRD
TERRITORIALITY
1067
mined in advance with time, territory, and observer selectedrandomly. For convenience, time
budget observations always started on the hour.
maintained by the station. All measurements,
except for rain amounts, were made continuously (24 hr/day) throughout the study.
NECTAR
STATISTICS
CONSUMPTION
To track patterns of food removal from the experimental territories nectar depletion from the
territorial feederswas recordedeach hour during
the hummingbirds’ active period. Changes in
feeder volumes were recorded to the nearest 0.2
ml. Syringe volumes were calibrated by measuring the mass of a volume of water inside the
syringe. Feederswere initially filled each day prior to the beginning of the active period, and the
final feeder measurement taken after the hummingbirds had gone to roost.
FEEDING
RATE
To estimate energy intake we measured feeding
rate for each hummingbird species.The feeder
used in these measurements was located outside
our laboratory window because the apparatus
could not be operated at territory sites.The feeders used, however, were identical to those in the
experimental territories. Feeding rate was determined by measuring the mass of feeder solution
removed by a hummingbird during a feedingbout
over time. Feeder mass was measured using a
calibrated strain gauge (Measurements Group,
Inc. EA-06- 125B2-350) attached to a brassbeam
from which the feederwas suspended.Mass measurements were accurate to 0.01 g. The strain
gaugewas calibrated by hangingprecisionweights
from the beam. Calibration was checked regularly to insure the accuracyof our measurements.
Output from the strain gaugewas sampled at 0.5
set intervals with a Campbell Scientific CR2 1X
data-logger.
TEMPERATURE
MEASUREMENTS
We monitored temperature each day of the study
in an unsheltered area near the feeding stations
3 m above the ground. Shaded ambient temperature (T,) was measured with a 24-gauge Cu-Cn
thermocouple and operative temperature (T,;
Winslow et al. 1937) a temperature increase or
decreasedue to radiative and convective factors,
with a copper sphere thermometer painted flatgray (Walsbergand Weathers 1986). Output from
the thermocouple and sphere thermometer was
sampled every minute and averaged every 15
min by a Campbell Scientific CR2 1X datalogger.
We measured precipitation with a rain gauge
Sample means were compared using paired and
unpaired Student-t tests when appropriate (Zar
1974). Comparison of morning, afternoon, and
evening samples was done using the KurskalWallis test (Zar 1974). Results are given as the
mean -t one standard deviation. Differences are
considered significant if P < 0.05.
RESULTS
WEATHER
Minimum T, occurred at 05:OOhr and averaged
9.8 f 2.O”C. Minimum T, also occurred at 05:
00 hr and was significantly lower than minimum
T, (t = 9.84, df = 19, P < 0.05) averaging 8.9 -t
2.8%. Maximum T, occurred at 15:00 hr and
averaged 28.0 f 5.9”C. Maximum T, also occurred at 15:00 hr and was significantly higher
than maximum T, (t = 23.5 1, df = 27, P < 0.05)
averaging 33.8 f 8.O”C.Precipitation during the
study totalled 5.55 cm, most of which occurred
on 1 June (0.74 cm), 10 June (2.41 cm), and 30
June (2.31 cm).
NON-AGGRESSIVE
BEHAVIOR
TERRITORIAL
Food availability did not appear to change the
basic characteristics of L. clemenciae territoriality, although the intensity and frequency of certain behaviors did vary (time budgets are summarized in Table 1; also, see below). Territory
owners spent significantly more time out-of-sight
when food was restricted than when food was
unlimited. Lampornis clemenciae spent the
greatestamount of time off their territory during
the afternoon period of the restricted food condition when feeders often were empty. During
this period, 00s time differed significantly from
00s time in the morning and evening (H =
15.322, df = 2, P < 0.05). Time-of-day did not
affect00s time when food was unlimited. When
afternoon measurements taken during the restricted food condition are eliminated from the
analysis, 00s time during the restricted period
is still significantly greater than when food is unlimited (t = 1.175, df = 85, P -C 0.05).
Territory owners were observed perching significantly longer when food was unlimited than
when food was restricted. When food was re-
1068
DONALD
TABLE
1.
R. POWERS
AND TODD
MCKEE
Time budget data for territorial Lampornisclemenciae.Values are expressed as min/hr +- 1 SD.
TCX-liiO~
Perching
Chasing
Feeding
Misc. Bight
00s
Unlimited
Restricted
17.82 + 8.00
9.05 + 10.83
2.992*
1.09 k 1.00
0.41 + 0.46
4.279*
1.04 f 0.83
0.58 + 0.58
3.181*
0.41 + 0.48
0.42 f 0.61
0.930
38.34 k 18.58
45.95 -c 13.70
2.356*
t
l
Indicates significant difference at
P < 0.05.
stricted time-of-day had a significant impact on
the time observed perching (H = 15.897, df = 2,
P < 0.05), with the highest amount of perching
observed in the morning and the lowest in the
afternoon. Perching time did not vary with timeof-day when food was unlimited.
Miscellaneous flight time did not vary between
conditions of food availability and averaged less
than 1% of the total time budget. However, miscellaneous flight varied significantly with timeof-day when food was restricted (H = 14.953, df
= 2, P < 0.05) whereas there was no time-ofday effect when food was unlimited.
ENERGY
INTAKE
When food was unlimited, an average of 10.6 +8.4 ml/hr (37.8 + 29.9 kJ/hr) of sucrosesolution
was consumed by hummingbirds throughout the
day (Fig. 1A). Total daily nectar removal from
the feedersaveraged 158.7 f 125.6 ml/day which
is equivalent to 484 Id/day. When food was restricted all 32.0 ml (114 kJ) of nectar placed in
the experimental feeders each day was consumed. Becauseof the manner in which nectar
was supplied, a bimodal feeding pattern was imposed (Fig. 1B).
The measured rate of food intake for Lampornis clemenciae was 2.22 g of sucrosesolution
per minute of feeding time. Territory owners fed
an average 1.04 min/hr (Table l), consuming an
estimated2.3 g/hr of sucrosesolution. Thus, when
food was unlimited, total daily energy intake is
estimated to be 113.8 + 91.3 Id/day (assuming
0.2 g sucrose/gsucrose solution). Although energy intake was high throughout the day, intake
did decline slightly as the day progressed(Fig.
2A). When food was restricted L. clemenciae’s
total daily energyintake decreasedto 63.8 f 63.8
k.I/day, with about 57% of their total energy consumption occurring in the morning (Fig. 2A).
Archilochus alexandri fed at a rate of 1.14 glmin
of sucrose solution for 1.8 min/hr, resulting in
an intake rate of 2.1 g/hr of sucrosesolution when
food was unlimited. This is equivalent to about
104.0 + 97.6 kJ/day. Like L. clemenciae, energy
intake for A. alexandri decreasedthroughout the
day when food was unlimited, although the decrease was sharper in A. alexandri (Fig. 2B).
However, A. alexandri still consumed large
amounts of nectar from territorial feeders
throughout the active period. When food was
restricted A. alexandri consumed only 27.6 +
50.2 M/day. Unlike L. clemenciae, the bulk of
A. alexandri’s energy intake from the experimental feeders, about 58%, occurred during the
evening (Fig. 2B).
Feeding bouts by hummingbirds were not always continuous. Often hummingbirds would
pauseone to several times during a single feeding
bout. The number of pausesoccurring during a
feeding bout did not vary significantly between
conditions of food availability. Lampornis clemenciae paused 0.5 + 1.3 times during a feeding
bout (range: 0 to 8 pausesper feeding bout). Archilochus alexandri paused 1.8 f 2.7 times per
feeding bout (range: 0 to 14 pauses).Pause rates
for the two speciesare significantly different (t =
10.15, df = 1,236, P < 0.05). There was no significant relationship between the number of
pausesoccurring during a feeding bout and the
amount of nectar consumed for either species.
The number of pauseswas therefore disregarded
in the calculation of feeding rate.
TERRITORIAL
AGGRESSION
Unlimited food. Intrusion rate by L. clemenciae
was 8.0 f 5.8 intrusions/hr and for A. alexandri
10.8 f 9.8 intrusions/hr. Lampornis clemenciae
intruders were chasedat a rate of 6.5 + 5.3 chases/hr (8 1%) whereasA. alexandri intruders were
chased at a rate of 1.2 + 2.4 chases/hr (11%).
Lampornis clemenciae intruders were able to feed
at territorial feeders 40% of the time whereas
76% of the A. alexandri intruders were able to
feed. The duration of feeding bouts for L. clemenciae and A. alexandri intruders was 0.07 +
1.1 set and 0.22 f 0.22 set, respectively.Feeding
bouts by intruding L. clemenciae were particu-
HUMMINGBIRD
TERRITORIALITY
1069
1300
1100
Hour
Feeders
filled
500
700
!300
1300
1100
1500
1700
1900
Hour
FIGURE 1. Number of times the territory owner visited the territorial feeder A) when food was unlimited
and B) when food was restricted. Each point representsthe number of feeder visits during an hour observation
period.
larly short because the territory owner immediately chasedthe intruder away from the feeder.
Average total time spent chasing intruders was
consistentbetween territories, but variability was
high over all time periods. ChasesofA. alexandri
intruders were usually short, lasting only a few
seconds,and terminated several meters from the
territorial feeder. Chases of conspecifics, however, could be one to two minutes and appeared
to be conducted at higher speedsand with greater
intensity than chasesof A. alexandri. Often L.
clemenciae other than the territory owner and
1070
DONALD R. POWERS AND TODD MCKEE
W
Unlimited
Restricted
A
Mornias
Evening
Afternoon
Time of Day
50
1
n
Unlimited
Restricted
B
Morning
Eveniag
Afternoon
Time of Day
FIGURE 2. The estimated amount of energyconsumedfrom territorial feedersby A) L. cIemenciueand B)
A. alexundri during the morning, midday, and evening. Estimation of energyconsumptionis based on measurementsof total feedingtime and feedingrate for each of the species.
intruder would participate in chasesof intruding
L. clemenciae.Some chasesinvolved as many as
four birds.
Restrictedfood. Lampornis clemenciae intrusion rate decreasedto 3.5 f 4.3 intrusions/hr
and A. alexandri intrusion rate dropped to 3.6
k 4.5 intrusions/hr. Both these values are significantly different from the unlimited food condition (t = 4.499, df = 101, P < 0.05 and t =
4.526, df = 101, P < 0.05 respectively). Territory
owners engagedin 2.8 f 3.9 chases/hr (80%) of
conspecificsand 1.7 f 2.6 (48%) chases/hrofA.
alexandri. Conspecific chaseswere significantly
less numerous than when food was unlimited (t
= 4.04 1, df = 10 1, P < 0.05), although territory
owners chasedthe same percentageof intruders.
The number of chasesof A. alexandri did not
differ from the unlimited food condition, but territory owners chased a higher proportion of intruders. The proportion of intruders gaining access to the territorial feeders decreased during
this portion of the experiment. Lampornis clemenciae intruders were able to feed only 29% of
the time whereasA. alexandri intruders fed only
58% of the time. The duration of intruder feeding
bouts was 0.06 f 1.0 set for Lampornis cle-
HUMMINGBIRD
TERRITORIALITY
1071
measures of daily energy expenditure to assess
energetic success.When food was unlimited L.
clemenciae territory owners consumed 114 kJ/
day, 39% more than their FMR measured with
doubly labeled water (82 kJ/day when food is
DISCUSSION
unlimited, Powers and Conley 1994). It appears
TERRITORIAL BEHAVIOR
therefore that L. clemenciae had no difficulty
Under both conditions of food availability ter- meeting its energetic needs assuming FMR was
ritory owners spent large amounts of time out- not substantially greater than 82 kJ/day. Yet,
of-sight (Table 1). BecauseL. clemenciaeestab- because most aspects of the environment are
lished territorial perchesin thick riparian habitat subjectto changefrom year-to-year, it is possible
it is likely that during most of the 00s time the that the higher energy intake indicates that FMR
in this study was higher than that measured by
territory owner was actually perched. Territory
owners usually went 00s during a chase and Powers and Conley (1994). Obvious factors that
could easily be missed returning to a perch high can affectFMR are the thermal environment and
in a tree. Often birds recordedas being 00s were differencesin activity. Maximum and minimum
detected when they began to call or sing. Thus, T, reported by Powers and Conley during their
we have likely underestimated perching time in PMR measurements were 34°C and 15”c, rethis study. A largeamount of perchingtime, about spectively. Although averagemaximum daytime
80% of the total activity budget, is characteristic T, in this study was the same as that reported by
of other hummingbirds that have been studied Powersand Conley, averageminimum nighttime
T, was 6°C cooler. If we assume that this 6°C
(e.g., Stiles 197 1).
Perches used by territory owners varied fre- temperature difference was maintained throughquently, even within a single day. No preference out the night, then FMR would be increased by
was shown for perchesnear the territorial feeder only 3 kJ (assumingthermal conductanceis 0.04
(within 5 m). Territory owners sang or called kJ hrr’“C-I; Iasiewski and Lasiewski 1967). It
regularly during each hour of observation while is therefore unlikely that differencesin the therperching (seeKuban and Neil1 1980, Powersand mal environment caused FMR to be substanConley 1994). During our observation periods tially different from that reported by Powers and
we could hear adjacent territory owners calling Conley. No time budgetswere conducted during
or singing simultaneously with the territory own- Powers and Conelv’s FMR measurements so it
er being observed. Kuban and Neil1 (1980) sug- is impossible to determine if a difference in acgestedthat the close proximity of calling L. cle- tivity level contributed to the high energy intake
observed in this study. However, food conmenciae males along with their complex
aggressivebehaviors might indicate lekking be- sumption in this study did not differ significantly
havior. The possibility of lek behavior, or some from measurements made under similar condiother structured social system, is certainly sup- tions by Powers and Conley (1994) suggesting
ported by the strong participation of territorial that energy requirements for territory owners in
males in aggressiveinteractions with conspecifics these two seasonswere comparable.
The feeding pattern exhibited when food is
regardlessof energy availability (seebelow). Flycatching and chaseactivity most often originated unlimited differed from that observed in other
from a territorial perch. These activities, al- free-living hummingbirds. For example, Anna’s
though significant to the territory owner, made Hummingbirds (Calypte anna) and Broad-tailed
Hummingbirds (Selasphorusplatycercus)exhibit
up only a fraction of an hourly time budget (Taa bimodal feeding pattern which includes a peak
ble 1).
in feeding activity prior to roosting (Wheeler
ENERGY INTAKE
1980, Calder et al. 1990). Calder et al. (1990)
Over the long-term hummingbirds, like any oth- suggestthat this feeding pattern is employed by
er animal, must consume an amount of food hummingbirds to delay the increasedcostof llight
sufficient to meet their daily energy needs. Alactivities that would result from mass gain asthough measurements of food consumption (en- sociatedwith the consumption of large volumes
ergy intake) are at best only rough estimates of of nectar. However, in this study feeding rate
FMR, they can be compared with more direct (visitsjhr) is relatively constant all day for both
menciae and 0.23 + 0.24 see for A. alexandri.
These values are not significantly different from
those measured when food was unlimited.
1072
DONALD
R. POWERS AND TODD
MCKEE
L. clemenciaeand A. alexandri (Fig. l), and energy intake &.I) declined continually, reaching its
lowest point just prior to the onset of the nocturnal fast (Fig. 2). Becausedaytime masschanges
of birds in this study were not monitored (e.g.,
Calder et al. 1990) we can only speculateas to
the reasons for this difference in foraging patterns. One possibility is that ifthe costof foraging
is low when food is unlimited, then birds that
synthesize the bulk of their fat in the morning
rather than in the evening might not be in danger
of experiencinga net energydeficit. This is mildly
supported by the fact that mass measurements
made throughout the day on L. clemenciaeand
A. alexandri when food is unlimited suggestthat
mass plateaus quickly in the morning and remains constant the rest of the day (Powers, unpubl. data).
INTRA-SPECIFIC
COMPETITION
Assuming the FMR of L. clemenciaeis about 82
Id/day, then to remain in energy balance they
must consume approximately an equivalent
amount of energy from their food source each
day. The ability of L. clemenciaeto intake enough
energy to maintain energy balance will depend
on the quality of their food source and on the
level of competition. In this study nearly all intra-specific interactions engaged in by territory
owners involved intruding male L. clemenciae.
On rare occasionsfemale L. clemenciae did attempt to feed at territorial feeders, but such intrusions were too infrequent to assessthe response of the territory owner. Our analysis
therefore focusessolely on male intruders.
The unlimited food availability condition of
this study presumably representsa high quality
food sourcethat is compact and easily defensible.
Under this condition intrusion rate was high,
averaging about 8 intruderslhr. Territory owners
were able to expel 8 1% of these intruders. Those
intruders that successfullyfed at the territorial
feeder usually did so while the territory owner
was away on a chase. This could be done easily
because intra-specific chases often last several
minutes and often involve up to four L. clemenciae. The reasonswhy so many male L. clemenciae were sometimes involved in territorial chases is unclear. However, it is possible that (1)
neighboring territory owners became involved
when a chase passed through their territory, or
(2) cooperation among neighboring territory
owners or intruders provides some energetic or
social benefit. Intra-specific chasesdid appear to
be more energetically expensive than inter-specific chases,at least in terms of their duration,
so cooperative defense might be beneficial to all
territory owners in a given area. Cooperative defensemight also make senseif L. clemenciaedoes
indeed form leks. Such behavioral mechanisms
would probably be more important when energy
is limited because when energy was unlimited
intake by territory owners exceeded their estimated FMR by 39%. Meeting daily energy demands was therefore not a problem for territory
owners even though numerous interactions took
place.
The 114 kJ/day provided to territory owners
during the food restriction experiment was presumed to represent a condition of energy limitation. This is supported by the fact that all food
provided during food restriction experiments was
consumed. Becausethis amount of energy is only
39% higher than the predicted L. clemenciae
FMR, the territory owner would presumablyhave
to work harder to retain 82 M for himself. This
task was made easier for the territory owner becauseintra-specific intrusions decreasedby 40%
when energy was limited, which should result in
an energy savings because there are fewer intruders to chase. This is supported by the fact
that energy consumption by territory owners decreased to 64 kJ/day. This is almost 30% less
than the DLW estimate of L. clemenciaeFMR.
Alternatively, territory owners might not have
been able to meet their energy needs from the
territorial feedersunder theseconditions and had
to forage elsewhereduring the afternoon period.
Extra-territorial foraging might have contributed
to the increased 00s time observed during food
restriction. Even if territory owners were unable
to meet their entire energy need from the territorial feeders, they still actively defended their
territories in the same manner for the duration
of the study. Chasesthat occurred were similar
to chasesobservedwhen food was unlimited and,
like the unlimited energy condition, the territory
owners chased80% of the total intra-specific intruders. Thus, regardless of the energy availability, territory owners worked vigorously to repel intra-specific intruders from their territories.
This suggeststhat aggressionservespurposesbeyond the simple defense of a food source,which
is consistent with our first hypothesis.
HUMMINGBIRD
TERRITORIALITY
1073
half the A. alexandri that intruded on their territory (although the total number of chasesdid
Male and female A. alexandri compete for many not differ between conditions). We believe this
of the same food resourcesas L. clemenciae at indicates that territory owners were more aware
the Southwestern Research Station in the Chir- of intruding A. alexandri. If correctthen this also
icahua Mountains. In most areas of its range, supports our first hypothesis. During this phase
male A. alexandri is a successfulterritorial spe- of the study, territory owners could not count on
cies (e.g., Ewald and Bransfield 1987). However, an endless supply of sucrose solution and preat our study site male A. alexandri behave as a sumably would incur cost if A. alexandri were
non-territorial species,presumably becausethey allowed to intrude unmolested as before. In adare unable to displace L. clemenciae, a species dition, because the number of intra-specific innearly three times larger. We have no quantita- teractionswas substantially decreasedduring this
tive data on the ratio of male to female A. al- experiment, territory owners would have more
exandri (not all birds were marked), but 70% of time to defend their feedersagainst inter-specific
the total intrusions on L. clemenciae territories intruders. This is supported by the fact that the
involved females. It is therefore likely that fe- proportion of successfulintrusions (where inmale A. alexandriare more numerous than males truders actually fed from the feeder) decreased
in our study area. We observed no discernible by 18% and the total number of successfulindifference in the interactions between territory trusionsdecreasedby 74%. Territory ownerswere
owners and intruding A. alexandri of different successfulin preventing substantial nectar loss
sexes,so the resultsof all A. alexandri intrusions to intruders, allowing A. alexandri to remove
are combined for analysis.
only 27 Id/day. (Wasps were not feeding on the
When food was unlimited, feeding from ex- sucrosesolution during this phase of the experperimental feedersdid not appear to be difficult iment.) This is enough energy to support only a
for A. alexandri. Territory owners chased only single A. alexandri based on the DLW estimate
11% of the A. alexandri intruders so most were of FMR (actually several A. alexandri got small
able to feed unmolested. The low proportion of amounts of sucrose solution). The I 14 Id/day
A. alexandri chased is consistent with observa- available should therefore have been enough to
tions made by Powers and Conley (1994) where meet a territory owner’s energy need even if their
L. clemenciae territory owners chased only 4%
FMR was not reduced when energy was restrictof the A. alexandri intruders. Potential reasons ed as suggestedabove, and assumingonly a small
for the low proportion of A. alexandri chased amount of nectar loss to intra-specific intruders
include: (1) becauseenergy was available in an (intruding L. clemenciae on average fed for less
unlimited amount, L. clemenciae could ignore than 1 sec/hr).
A. alexandri without cost, (2) A. alexandri often
An alternative explanation for the behavioral
did not approach the experimental feeders until
changesthat occurred when food was restricted
the territory owner was off the territory, thus is that the data were collected during a period
reducing the chance of being chased,and (3) inwhen the birds were shifting from breeding to
tra-specific interactions were relatively frequent non-breeding status. The hummingbirds arrive
and received higher priority from territory own- in the Chiricahuas in late March to early April,
ers. Additional evidence that A. alexandri was and breed (in some cases) until mid July (see
able to foragerelatively freely from experimental Johnsgard1983 for discussion).During the latter
feederswhen food was unlimited comesfrom the portion of June (when food restriction experifact that 484 Id/day was being removed from the ments were conducted)it is possiblethat the birds’
feeders. This is slightly lessthan six times a ter- behavioral pattern changednaturally as our study
ritory owner’s FMR based on DLW. Substantial encroached on the end of the breeding season.
amounts of sucrosesolution were therefore being However, when the behavior of hummingbirds
removed by intruders, predominantly A. alexin this study during the unlimited food condition
andri (about 103 M/day) and wasps.
(observed in early June) is compared to that of
During the food restriction experiment the the same hummingbird population in early July
number of intruding A. alexandri was greatly re- under the same experimental conditions (Powers
duced. However, territory owners chasednearly and Conley 1994), no discernible difference is
INTER-SPECIFIC
COMPETITION
1074
DONALD R. POWERS ANDTODD MCKEE
observed. In addition, other studies involving
these hummingbird populations (Pimm 1978,
Pimm et al. 1985) report data collected between
April and June, and do not identify behavioral
shifts attributable to seasonal change. We therefore feel that it is unlikely that the time of measurement in this study affect our behavioral data,
and that the changes observed were due solely
to experimental manipulation.
CONCLUSIONS
This study illustrates behavioral changes that occur in territorial L. clemenciae in response to
changes in food availability. When food availability is high L. clemenciaeappears to focus on
social interaction with conspecifics, while ignoring frequent visits by inter-specific intruders (A.
alexandrz)that can remove large amounts of food
from the territorial food source. Lampornis clemenciae does not seriously exclude A. alexandri
until food becomes limited. This suggests aggression towards intra-specific and inter-specific
intruders is initiated based on different primary
factors. Inter-specific aggression appears motivated primarily by food availability. If L. clemenciae habitat has a rich supply of food then
overall hummingbird diversity is likely to be high
because other species will have access to food
sources, even if they are being defended.
Although food availability certainly plays a role
in initiating aggression against conspecific intruders, social factors are likely just as important.
Territory owners exhibited strong aggression
against conspecific intruders under all conditions
of food availability. The only aspect of this aggression that was clearly related to changes in
food availability was frequency of occurrence. It
makes sense that when food availability is low
that even L. clemenciae density will be decreased. However, the characteristics of aggressive interactions between territory owners and
intruders were not obviously altered when food
availability was changed. Thus, the territorial behavior exhibited by L. clemenciaepossibly serves
a function beyond the simple defense of a food
source (hypothesis #l). We must, however, be
cautious in our interpretations and recognize the
possible alternatives. For instance, the food source
used in this study was artificial. It is therefore
possible that the emphasis on conspecific chases
have been selected for evolutionarily by natural
conditions that depart widely from artificial feeders. A second alternative results from the like-
lihood that the larger L. clemenciae intruders
would, over the long term, consume more nectar
than the smaller A. alexandri. For this reason
alone it might be more important for territory
owners to focus on inter-specific intruders.
A second conclusion from this study centers
around the behavior ofA. alexandri. A. alexandri
typically is a territorial species. However, at our
study site they were non-territorial. This suggests
that the territorial nature of A. alexandri is not
a requirement for survival. This observation
brings into question the notion that territoriality,
at least in A. alexandri, is an evolutionarily derived trait rather than one of perhaps many behavioral options that can be employed in a variety of community interactions.
ACKNOWLEDGMENTS
We are gratefulto Wesley Weathers,Marcus Webster,
Harry Tiebout III, and two anonymousreviewersfor
commenting on an earlier version of this manuscript.
Wesley Weathers, Wade Sherbrook, Rim Sullivan,
Aaron Irish, and Scott Smithson provided valuable
technicalassistance.We thank the American Museum
of Natural History for useof the facilitiesat the SouthwesternResearchStation. This project was supported
by a National GeographicSocietyGrant-in-aid-of-Research(#4458-9 1).
LITERATURE
CITED
D. P. 1987. Economicsof breedingterritoriality in male Calliope Hummingbirds. Auk
104242-253.
BAKKEN,G. S. 1976. A heat transferanalysisof animals: unifying conceptsand the application of
metabolismchamberdata to field ecology.J. Theoretical Biology 60:337-384.
BROWN,
J. L. 1964. The evolution of diversity in
avian territorial systems.Wilson Bull. 76:160-169.
CALDER,W. A., L. L. CALD~~, AND T. D. FRAIZER.
1990. The hummingbird’s restraint: a natural
modelforweightcontrol.Experientia46:999-1002.
CARPENTER,
F. L., AND R. E. MACMILLEN. 1976.
Threshold model of feedingterritoriality and test
with a Hawaiian Honeycreeper.Science194:639642.
EwALD,P.W.,ANDR.J. B~ANSFIFXD.1987. Territory
quality and territorial behavior in two sympatric
speciesof hummingbirds.BehavioralEcology
_. and
Sociobiology20~285-293.
EWALD.P. W.. AND F. L. CARPENTER.1978. Territorial responsesto energy manipulations in the
Anna Hummingbird. Oecologia3 1:277-292.
FEINSINGER,
P. 1986. Organization of a tropical guild
of nectarivorousbirds. EcologicalMonographs46:
257-291.
Fa~~r+nrrrr,C. R. 1983. Cost of aggressionin trout
and pupfish, p. 117-l 38. In W. P. Aspey and S.
L. Lustick [eds.], Behavioral energetics:survival
in vertebrates.Ohio StateUniv. Press,Columbus.
ARMSTRONG,
HUMMINGBIRD
JOHNSGARD,
P. A. 1983. The hummingbirdsofNorth
America. Smithsonian Institution Press, Washington, DC.
KODRIC-BROWN,
A., ANDJ. H. BROWN. 1978. Influence of economics,interspecificcompetition, and
sexual dimorphism on territoriality of migrant
Rufous Hummingbirds. Ecoloav 59:285-296.
KUBAN,J. F., ANDR. c NEILL. 198E Feedingecology
of hummingbirds in the highlands of the Chisos
Mountains, Texas. Condor 82: 180-l 85.
LASIEWSKI,
R. C., AND R. J. LAsIEwsKI. 1967. Physiological responsesof the Blue-throated and Rivoli’s Hummingbirds. Auk 84:34-48.
P~MM,S. L. 1978. An experimental approach to the
effects of predictability on community structure.
American Zoologist 18:797-808.
PIMM, S. L., M. L. RosENZWEIG,AND W. MITCHELL.
1985. Competition and food selection:field tests
of a theory. Ecology 66:798-807.
POWERS,
D. R. 1987. Effectsof variation in food quality on the breedingterritoriality ofthe male Anna’s
Hummingbird. Condor 89:103-l 11.
POWERS,
D. R., ANDT. M. C~NLEX. 1994. Field metabolic rate and food consumption of two sympattic hummingbird speciesin southeasternArizona.
Condor 96:141-150.
TERRITORIALITY
1075
STILES,F. G. 1971. Time, energy, and territoriality
of the Anna Hummingbird (Culypte anna). Science 173:818-820.
S-mm, F. G. 1975. Ecology,floweringphenology,and
humminabird uollination of some Costa Rican
Heliconia species.Ecology 56:285-30 1.
WAISBJZRG,G. E., AND W. W. Wtr~rnnas. 1986. A
simple technique for estimating operative environmental temperature. J. Thermal Biology 11:
67-72.
WFXX, R. C., M. J. A~TLE,AND W. H. Brrvaa [EDS.]
1983. CRC handbook of chemistry and physics.
_
64th ed. CRC Press,Boca Raton, FL.
WHEELER.T. G. 1980. Exneriments in feedine behavior of the Anna Hummingbird. WilsonBull.
92:53-62.
Wnusrow, C. E., L. P. HE~UUNGTON,
ANDA. P. GAGGE.
1937. Physiologicalreactionsof the human body
to varying environmental temperatures. Am. J.
Physiology 120~1-22.
WOLF,L. L. 1978. Aggressivesocial organization in
nectarivorousbirds. American Zoologist 18:765778.
ZAR,J. H. 1974. Biostatisticalanalysis.PrenticeHall,
NJ.